Anne PCW16

[steve]

Instead of arguing over trivialities, you could explain how you would get the Z80 to run at the same speed as it would in a CPC.

[TFM]

Instead of arguing over trivialities, you could explain how you would get the Z80 to run at the same speed as it would in a CPC.

We should continue this discussion in the CPC# Thread. Let's stay PCW16 here

[MaV]

Thread necromancy:

@TFM:

According to these sites, the PCW16 uses a standard three-button mouse with the "Mouse Systems" protocol.
Amstrad PcW16

JOYCE: Downloads
In the file bin/AnneSerialMouse.cxx there's a reference to Mouse Systems.

And finally here are the "Mouse Systems" protocol specifics:
What protocol do mice use?

In the end, you'll need to find a 3-button serial mouse, as the Mouse Systems protocol seems to be standard for 3-button PC mice.


MaV

[ralferoo]

Hehe! The British won't take DIN, but DIN would be good for the whole world ;-)

Ummm, what? Plenty of stuff had DIN connectors in the 80s in Britain as the default multi-pole connector choice. And we used to refer to them all as n-pin DIN too (although I can't remember how we used to differentiate between the 180° and 270° versions).

Reel-to-reel tape recorders, other tape recorders, CPC monitor, Dragon joystick ports, etc. You can still just walk into any branch of Maplin and buy them ready for soldering on to things...

[Bryce]

Even the "As British as it gets" BBC computers used DIN connectors!

Edit: I probably should mention, there's an entire PCW16 including mouse/manuals and all the rest on offer over on Amibay at the moment: http://www.amibay.com/showthread.php?t=46396

Bryce.

[robcfg]

Are the protocols that different so that one has tried to build a PS/2 or USB to mouse systems adapter?

I'm pretty sure that it would be cheaper than trying to find that kind of mice.


Edit: I found this description of the protocols.

[TFM]

DIN in GB? I never stop learning!

Thanks Rob for the link!

[MaV]

Well, yeah, someone is making money out of this market niche:
PS/2 and Serial Mouse Protocol Converters from Vetra Systems, the Industry's

VIP-327-PS-3B for 65$

Perhaps there's a Chinese company for that.

They may be hard to find, but buying a used serial three button mice will be much cheaper, even if you have to wait a little to find one.

Anyway, as I said in my previous post, the point is: The PCW16's mouse is a PC standard 3 button mouse and has a known protocol.

[TFM]

Yes, and it's extremely hard to get it. I tried about 10 mice, and not a single one worked.

[Bryce]

Here's the original user manual on offer, in case you might get more info there: Amstrad PCW16 User Manual 1996 | eBay

Bryce.

[TFM]

Sadly only UK 

[Prodatron]

Exactly! And this ASIC lacks some of the illegal Z80 instructions. (I'm not talking about the undocumented thought, that's a difference).

Sorry for digging out this old thread, but this is exactly, what I am wondering about:
Does the "Z80" in the PcW16 ANNE Asic support all kind of 8bit index register access? Like LD r,IXL or DEC IYH or ADD IXH etc. etc.

CU,
Prodatron

[robcfg]

I've just asked Habi (CP/M Box author) and he thinks that all these are supported . He also thinks it passes all zexall test. Not so sure it supports the illegal instructions. He'll be testing it and I'll be telling you the results as soon as he tells me.

[Prodatron]

Thanks a lot!

[TFM]

Sorry for digging out this old thread, but this is exactly, what I am wondering about:
Does the "Z80" in the PcW16 ANNE Asic support all kind of 8bit index register access? Like LD r,IXL or DEC IYH or ADD IXH etc. etc.

CU,
Prodatron

Looked at the PCB (back the day), it's a kind of custom chip instead of the Z80. As I know it supports at least the commonly used undocumented and illegal instructions. But honestly I didn't make a full test, just used couple source examples.

[robcfg]

Habi just confirmed that the pcw16 passes all zexall tests, and at 4x speed!

He told me that scf/ccf don't work because it's a nec core, but it's normal.

He thinks that the undocumented/illegal instructions should work fine as with zilog's z80.

[Prodatron]

Thanks a lot for the info!
SCF, CCF don't work?? Phew, but these are 100% "official" Z80 commands and sometimes quite usefull. Or do they still set/negate the CF but don't modify N,H?

[Habi]

Both SCF and CCF work perfectly; it's just that the undocumented flags behave in a different way.

I'll try to test all the undocumented instructions (including duplicates and double nops) and illegal ones (with redundant or mutual exclusive prefixes), just to be sure.

[Prodatron]

Thanks, Habi, and good to read from you here! 

[TFM]

Both SCF and CCF work perfectly; it's just that the undocumented flags behave in a different way.

I'll try to test all the undocumented instructions (including duplicates and double nops) and illegal ones (with redundant or mutual exclusive prefixes), just to be sure.

That's awesome! And welcome here, even if my welcome comes a bit late. 

[Habi]

Well, it appears to behave exactly like a Zilog Z80.

Double o exclusive prefixes doesn't cancel each other, redundant do nothing. The double nops ($ED) behave like that, and also the undocumented opcodes worked, redundant (NEG, IM x, ...) or not (SLL).

There are two things I would like to test also: the OUT (C), 0 opcode (I suspect will be more like OUT (C), $FF because it's CMOS) and MemPTR (which I believe will be again the same as Zilog).

I'll let you know when I have more information.

[Executioner]

Both SCF and CCF work perfectly; it's just that the undocumented flags behave in a different way.

Zexall doesn't test the SCF/CCF flags properly, and I don't think it's been mentioned here before. I have the info on it and it's something I'm planning on implementing in my emulators for the next release.

[Habi]

Finally I've had some time to test these things.

The Out puts an $FF on the bus (as expected), and MemPTR works the same way as a Zilog processor.

The PCW ASIC has also some kind of buffer for I/O which can be made float to any value, but that's another story...

Zexall doesn't test the SCF/CCF flags properly, and I don't think it's been mentioned here before. I have the info on it and it's something I'm planning on implementing in my emulators for the next release.

Yes, I know, but it is good enough to differentiate between Zilog and Nec, which was my point being it a NEC macrocell.

And for the record: there is NO good SCF / CCF test, because it's dependant on the previous instruction, as seen on [Z80 at Scratchpad, the home of unlimited fan-fiction mini-wikis!].

This article is particularly interesting: [Down to the silicon: how the Z80's registers are implemented]. It appears flags are latched in a temporal register, so we have here something similar to MemPTR logic.

For example, my Z80 emulation passes all the exhaustive Spectrum tests (way better than ZexAll) except the one specifically made to expose this double behaviour of SCF and CCF. I've some information on it also, but it's a bit confusing; so I'm waiting to someone translate those high resolution z80 scans into transistors.

Edit: I wold like to clarify that I was referring the NEC (uPD780) version of Z80. The original Zilog behaviour is known thanks to Patrik Rak, and of course it's correctly handled in my Z80 emulation.

[Prodatron]

As long as SCF, CCF and all 8bit IX/IY Index register commands (with IXL, IXH, IYL, IYH) are working "normal" I am very happy! 

[zhulien]

Hi, an old thread...


but... has anyone done a dump and full disassembly of the OS yet?  What is the memory map like on the PCW16?  Could the OS be modified to run on CPC?